vllm: [Bug]: cross-node tensor parallel + cudagraph issue
Your current environment
The output of `python collect_env.py`
Collecting environment information...
PyTorch version: 2.5.1+cu118
Is debug build: False
CUDA used to build PyTorch: 11.8
ROCM used to build PyTorch: N/A
OS: CentOS Linux 7 (Core) (x86_64)
GCC version: (GCC) 10.2.0
Clang version: Could not collect
CMake version: version 3.22.1
Libc version: glibc-2.17
Python version: 3.10.4 (main, Mar 31 2022, 08:41:55) [GCC 7.5.0] (64-bit runtime)
Python platform: Linux-3.10.0-957.el7.x86_64-x86_64-with-glibc2.17
Is CUDA available: True
CUDA runtime version: 12.1.66
CUDA_MODULE_LOADING set to: LAZY
GPU models and configuration:
GPU 0: NVIDIA A800-SXM4-80GB
GPU 1: NVIDIA A800-SXM4-80GB
GPU 2: NVIDIA A800-SXM4-80GB
GPU 3: NVIDIA A800-SXM4-80GB
GPU 4: NVIDIA A800-SXM4-80GB
GPU 5: NVIDIA A800-SXM4-80GB
GPU 6: NVIDIA A800-SXM4-80GB
GPU 7: NVIDIA A800-SXM4-80GB
Nvidia driver version: 535.104.05
cuDNN version: Could not collect
HIP runtime version: N/A
MIOpen runtime version: N/A
Is XNNPACK available: True
CPU:
Architecture: x86_64
CPU op-mode(s): 32-bit, 64-bit
Byte Order: Little Endian
CPU(s): 128
On-line CPU(s) list: 0-127
Thread(s) per core: 2
Core(s) per socket: 32
Socket(s): 2
NUMA node(s): 2
Vendor ID: GenuineIntel
CPU family: 6
Model: 106
Model name: Intel(R) Xeon(R) Platinum 8358P CPU @ 2.60GHz
Stepping: 6
CPU MHz: 3199.853
CPU max MHz: 3400.0000
CPU min MHz: 800.0000
BogoMIPS: 5200.00
Virtualization: VT-x
L1d cache: 48K
L1i cache: 32K
L2 cache: 1280K
L3 cache: 49152K
NUMA node0 CPU(s): 0-31,64-95
NUMA node1 CPU(s): 32-63,96-127
Flags: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb rdtscp lm constant_tsc art arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc aperfmperf eagerfpu pni pclmulqdq dtes64 monitor ds_cpl vmx smx est tm2 ssse3 sdbg fma cx16 xtpr pdcm pcid dca sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand lahf_lm abm 3dnowprefetch epb cat_l3 intel_pt ssbd mba ibrs ibpb stibp ibrs_enhanced tpr_shadow vnmi flexpriority ept vpid fsgsbase tsc_adjust bmi1 hle avx2 smep bmi2 erms invpcid rtm cqm rdt_a avx512f avx512dq rdseed adx smap avx512ifma clflushopt clwb avx512cd sha_ni avx512bw avx512vl xsaveopt xsavec xgetbv1 cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local dtherm ida arat pln pts hwp hwp_act_window hwp_epp hwp_pkg_req avx512vbmi umip pku ospke avx512_vbmi2 gfni vaes vpclmulqdq avx512_vnni avx512_bitalg avx512_vpopcntdq spec_ctrl intel_stibp flush_l1d arch_capabilities
Versions of relevant libraries:
[pip3] numpy==1.26.3
[pip3] nvidia-cublas-cu11==11.11.3.6
[pip3] nvidia-cublas-cu12==12.4.5.8
[pip3] nvidia-cuda-cupti-cu11==11.8.87
[pip3] nvidia-cuda-cupti-cu12==12.4.127
[pip3] nvidia-cuda-nvrtc-cu11==11.8.89
[pip3] nvidia-cuda-nvrtc-cu12==12.4.127
[pip3] nvidia-cuda-runtime-cu11==11.8.89
[pip3] nvidia-cuda-runtime-cu12==12.4.127
[pip3] nvidia-cudnn-cu11==9.1.0.70
[pip3] nvidia-cudnn-cu12==9.1.0.70
[pip3] nvidia-cufft-cu11==10.9.0.58
[pip3] nvidia-cufft-cu12==11.2.1.3
[pip3] nvidia-curand-cu11==10.3.0.86
[pip3] nvidia-curand-cu12==10.3.5.147
[pip3] nvidia-cusolver-cu11==11.4.1.48
[pip3] nvidia-cusolver-cu12==11.6.1.9
[pip3] nvidia-cusparse-cu11==11.7.5.86
[pip3] nvidia-cusparse-cu12==12.3.1.170
[pip3] nvidia-ml-py==12.560.30
[pip3] nvidia-nccl-cu11==2.21.5
[pip3] nvidia-nccl-cu12==2.21.5
[pip3] nvidia-nvjitlink-cu12==12.4.127
[pip3] nvidia-nvtx-cu11==11.8.86
[pip3] nvidia-nvtx-cu12==12.4.127
[pip3] pynvml==11.5.3
[pip3] pyzmq==26.2.0
[pip3] torch==2.5.1+cu118
[pip3] torch-tb-profiler==0.4.3
[pip3] torchaudio==2.5.1+cu118
[pip3] torchdata==0.9.0
[pip3] torchmetrics==1.6.0
[pip3] torchvision==0.20.1+cu118
[pip3] transformers==4.48.2
[pip3] transformers-stream-generator==0.0.5
[pip3] triton==3.1.0
[pip3] triton-nightly==3.0.0.post20240422051420
[conda] numpy 1.26.3 pypi_0 pypi
[conda] nvidia-cublas-cu11 11.11.3.6 pypi_0 pypi
[conda] nvidia-cublas-cu12 12.4.5.8 pypi_0 pypi
[conda] nvidia-cuda-cupti-cu11 11.8.87 pypi_0 pypi
[conda] nvidia-cuda-cupti-cu12 12.4.127 pypi_0 pypi
[conda] nvidia-cuda-nvrtc-cu11 11.8.89 pypi_0 pypi
[conda] nvidia-cuda-nvrtc-cu12 12.4.127 pypi_0 pypi
[conda] nvidia-cuda-runtime-cu11 11.8.89 pypi_0 pypi
[conda] nvidia-cuda-runtime-cu12 12.4.127 pypi_0 pypi
[conda] nvidia-cudnn-cu11 9.1.0.70 pypi_0 pypi
[conda] nvidia-cudnn-cu12 9.1.0.70 pypi_0 pypi
[conda] nvidia-cufft-cu11 10.9.0.58 pypi_0 pypi
[conda] nvidia-cufft-cu12 11.2.1.3 pypi_0 pypi
[conda] nvidia-curand-cu11 10.3.0.86 pypi_0 pypi
[conda] nvidia-curand-cu12 10.3.5.147 pypi_0 pypi
[conda] nvidia-cusolver-cu11 11.4.1.48 pypi_0 pypi
[conda] nvidia-cusolver-cu12 11.6.1.9 pypi_0 pypi
[conda] nvidia-cusparse-cu11 11.7.5.86 pypi_0 pypi
[conda] nvidia-cusparse-cu12 12.3.1.170 pypi_0 pypi
[conda] nvidia-ml-py 12.560.30 pypi_0 pypi
[conda] nvidia-nccl-cu11 2.21.5 pypi_0 pypi
[conda] nvidia-nccl-cu12 2.21.5 pypi_0 pypi
[conda] nvidia-nvjitlink-cu12 12.4.127 pypi_0 pypi
[conda] nvidia-nvtx-cu11 11.8.86 pypi_0 pypi
[conda] nvidia-nvtx-cu12 12.4.127 pypi_0 pypi
[conda] pynvml 11.5.3 pypi_0 pypi
[conda] pyzmq 26.2.0 pypi_0 pypi
[conda] torch 2.5.1+cu118 pypi_0 pypi
[conda] torchaudio 2.5.1+cu118 pypi_0 pypi
[conda] torchdata 0.9.0 pypi_0 pypi
[conda] torchmetrics 1.6.0 pypi_0 pypi
[conda] torchvision 0.20.1+cu118 pypi_0 pypi
[conda] transformers 4.48.2 pypi_0 pypi
[conda] transformers-stream-generator 0.0.5 pypi_0 pypi
[conda] triton 3.1.0 pypi_0 pypi
[conda] triton-nightly 3.0.0.post20240422051420 pypi_0 pypi
ROCM Version: Could not collect
Neuron SDK Version: N/A
vLLM Version: 0.7.3.dev90+gd59def47
vLLM Build Flags:
CUDA Archs: Not Set; ROCm: Disabled; Neuron: Disabled
GPU Topology:
GPU0 GPU1 GPU2 GPU3 GPU4 GPU5 GPU6 GPU7 NIC0 NIC1 NIC2 NIC3 NIC4 NIC5 CPU Affinity NUMA Affinity GPU NUMA ID
GPU0 X NV8 NV8 NV8 NV8 NV8 NV8 NV8 PXB NODE SYS SYS NODE PXB 0-31,64-95 0 N/A
GPU1 NV8 X NV8 NV8 NV8 NV8 NV8 NV8 PXB NODE SYS SYS NODE PXB 0-31,64-95 0 N/A
GPU2 NV8 NV8 X NV8 NV8 NV8 NV8 NV8 NODE PXB SYS SYS NODE NODE 0-31,64-95 0 N/A
GPU3 NV8 NV8 NV8 X NV8 NV8 NV8 NV8 NODE PXB SYS SYS NODE NODE 0-31,64-95 0 N/A
GPU4 NV8 NV8 NV8 NV8 X NV8 NV8 NV8 SYS SYS PXB NODE SYS SYS 32-63,96-127 1 N/A
GPU5 NV8 NV8 NV8 NV8 NV8 X NV8 NV8 SYS SYS PXB NODE SYS SYS 32-63,96-127 1 N/A
GPU6 NV8 NV8 NV8 NV8 NV8 NV8 X NV8 SYS SYS NODE PXB SYS SYS 32-63,96-127 1 N/A
GPU7 NV8 NV8 NV8 NV8 NV8 NV8 NV8 X SYS SYS NODE PXB SYS SYS 32-63,96-127 1 N/A
NIC0 PXB PXB NODE NODE SYS SYS SYS SYS X NODE SYS SYS NODE PIX
NIC1 NODE NODE PXB PXB SYS SYS SYS SYS NODE X SYS SYS NODE NODE
NIC2 SYS SYS SYS SYS PXB PXB NODE NODE SYS SYS X NODE SYS SYS
NIC3 SYS SYS SYS SYS NODE NODE PXB PXB SYS SYS NODE X SYS SYS
NIC4 NODE NODE NODE NODE SYS SYS SYS SYS NODE NODE SYS SYS X NODE
NIC5 PXB PXB NODE NODE SYS SYS SYS SYS PIX NODE SYS SYS NODE X
Legend:
X = Self
SYS = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI)
NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node
PHB = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU)
PXB = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge)
PIX = Connection traversing at most a single PCIe bridge
NV# = Connection traversing a bonded set of # NVLinks
NIC Legend:
NIC0: mlx5_2
NIC1: mlx5_3
NIC2: mlx5_4
NIC3: mlx5_5
NIC4: mlx5_bond_0
NIC5: mlx5_bond_1
🐛 Describe the bug
when I set tp_size=16, it occurs the following bugs. However, when tp_size <= 8, everything is fine.
Before submitting a new issue…
- Make sure you already searched for relevant issues, and asked the chatbot living at the bottom right corner of the documentation page, which can answer lots of frequently asked questions.
About this issue
- Original URL
- State: closed
- Created 2 years ago
- Comments: 18 (5 by maintainers)
some people tried
pip install -U nvidia-nccl-cu12and solved some NCCL-related bugs. I’m not sure if it works for you.Update: When I set tensor_parallel_size=16 in offline inference, and I set enforce_eager=True, it can work. However, when I set enforce_eager=False(default), the above problem still occurs. So I think maybe the cuda graph is not compatible with tp_size > 8?